What Does “Know the Enemy and Know Yourself” Actually Promise?

A fresh translation

Original text

Words and concepts

• 彼 (bi)
  “The other” or “the opposing side.” “Enemy” is natural in context, but the pronoun itself names the counterpart rather than defining a personal foe.
• 己 (ji)
  Oneself. In this chapter that includes the condition of one’s forces, command, numbers, preparedness, and relation between ruler and general.
• 知 (zhi)
  To know or understand. The preceding list makes this practical discrimination: knowing when battle is possible, how numbers can be used, and whether preparations are complete.
• 殆 (dai)
  Danger, jeopardy, or a precarious position. It does not mean “never lose.”
• 百戰
  “A hundred battles,” a conventional expression for repeated engagements, not a statistical sample of exactly one hundred.

The famous quotation is only the first case

“Know your enemy and know yourself” is often detached from what follows. On its own, it can sound like a complete theory: gather information about both sides and victory is assured. The received chapter does not end after that memorable pair.

It immediately supplies two controls. Knowledge of oneself without knowledge of the other produces mixed results. Knowledge of neither produces danger every time. The three clauses form a descending scale, not three unrelated sayings.

That scale matters because the first result is commonly overstated. The text says “not in peril,” using dai. It does not say “win every battle.” The distinction leaves room for refusal, delay, maneuver, and the judgment that battle should not be offered at all.

The promise concerns exposure to danger. A commander who understands both sides can avoid entering contests whose conditions make disaster likely. Victory may follow, but the clause is phrased as freedom from jeopardy rather than an endless winning record.

What must be known?

The quotation itself does not provide a checklist. Its position does. Immediately before it, the chapter names five ways to recognize victory: knowing when one can and cannot fight; understanding the use of many and few; having superiors and subordinates share the same aim; waiting prepared for the unprepared; and having a capable general whom the ruler does not restrain.

These are not items of secret intelligence about an opponent alone. Several concern one’s own side. Can the force act? Can numbers be used effectively? Is command coherent? Has preparation been completed? Does political interference obstruct the general?

“Know yourself” therefore reaches beyond confidence or personal character. The “self” is an organized force with limits, relationships, and material conditions. A leader who knows private intentions but misreads troops, authority, or readiness does not satisfy the chapter’s test.

“Know the other” is equally concrete. It requires comparison. Numbers mean little without knowing how they can be employed. Readiness matters because another side may be unready. A decision to fight becomes intelligible only in relation to an opposing condition.

Why one victory and one defeat?

The middle clause is unusually balanced: one victory, one defeat. It does not offer a precise probability. Classical parallel phrasing compresses an unstable pattern into a memorable pair.

Self-knowledge supplies something real. A force that understands its own capacities is not wholly blind. It may choose a suitable formation, protect a weakness, or decline an impossible task. That is why the middle condition is not identical to knowing neither side.

Yet the missing half prevents reliable judgment. An action suited to one’s own strength may meet a prepared opponent, unfavorable numbers, or a plan already anticipated. Good internal assessment cannot decide a relational question by itself.

The formula thus rejects inward certainty as a substitute for inquiry. Discipline, courage, and clarity within one’s own camp are valuable, but the result still varies if the counterpart remains unknown.

The chapter begins by preserving, not destroying

The title “Planning the Attack” can invite a narrow expectation of offensive technique. The chapter begins elsewhere. Taking an opposing state intact is better than breaking it; preserving an army is better than destroying it; winning a hundred battles is not the highest excellence; causing an opposing force to submit without battle is better.

It then ranks targets of action. Plans come first, alliances next, armies after that, and fortified cities last. Siege is described as a last resort because preparations consume months and an angry commander may sacrifice a third of the troops without taking the walls.

Against that background, “not in peril” is not weak language. It belongs to a chapter that measures skill partly by what need not be destroyed. Knowledge does not merely improve performance once battle begins. It helps determine whether battle, siege, delay, or another approach is warranted.

The maxim is consequently less theatrical than its reputation. Its surrounding argument favors wholeness, economy, and discrimination over repeated displays of battlefield success.

Knowing oneself includes limits on the ruler

Before the five conditions, the chapter lists three ways a ruler can trouble an army. He may order advance or retreat without knowing whether either is possible. He may govern military affairs as though they were ordinary civil administration. He may assign responsibility without understanding military authority.

The result is confusion and doubt. Victory is drawn away by disorder created from within. This warning changes the sense of “self.” The danger is not only a commander’s miscalculation. The ruler-general relationship belongs inside the diagnosis.

One of the five conditions makes that point explicit: a capable general wins when the ruler does not obstruct him. Knowing oneself therefore includes knowing where authority should stop. The chapter does not celebrate unlimited autonomy; it identifies mismatched intervention as a source of operational ignorance.

This also explains why the saying cannot be reduced to espionage. Information about the other side cannot repair a command system that refuses to understand its own competence and boundaries.

A textual variant sharpens the reading

The source transcription used here ends the third clause with 殆, “be in peril,” matching the first clause’s result. Other transmitted versions read 敗, “be defeated,” in that final position.

The variant does not erase the three-part structure, but it affects the symmetry. With 殆, the first and third cases answer each other exactly: knowledge of both means no peril; knowledge of neither means certain peril. With 敗, the last clause states the realized outcome of defeat.

An article should not silently combine those readings. The quoted original retains 殆 because that is the wording in the designated source text. The existence of 敗 is noted because readers may encounter it in modern databases and translations.

Neither reading turns the first clause into “you will win one hundred times.” The verb attached to complete knowledge remains a statement about peril.

What these three clauses cannot establish

They do not establish that sufficient information eliminates uncertainty. Information can be false, late, partial, or misunderstood. The passage states a standard of knowledge, not a procedure that guarantees omniscience.

They do not prove that every battle should be fought once both sides are understood. The preceding five conditions begin with knowing when one can and cannot fight. Knowledge may lead to abstention.

They do not reduce the Sunzi to a manual of personal self-awareness. The relevant “self” includes forces, numbers, readiness, command, and political relations. Modern psychological applications are later analogies, not the historical scope of the clause.

Finally, they do not record a verified utterance by a single historical author at a known moment. The work is traditionally attributed to Sun Wu, while the formation and transmission of the received thirteen chapters remain subjects of scholarship. The secure object here is the wording and placement of the received text.

Read the whole triad

The compactness of the saying has helped it travel. It has also made the first eight characters easy to separate from the argument that controls them.

Read as a triad, the passage becomes more demanding. Knowing oneself is necessary but incomplete. Knowing the other without understanding one’s own system is not even granted a separate category. Knowing both does not promise spectacle; it promises release from repeated peril.

The closing line of “Planning the Attack” therefore returns the reader to judgment. Skill lies not in claiming certainty, nor in fighting until success appears, but in distinguishing conditions before committing a force. That is why the chapter can praise victory without battle and end with a sentence so often mistaken for a guarantee of victory.

Source

Sunzi (孫子), chapter 3, “Planning the Attack” (謀攻), final three clauses. The characters were checked against the chapter transcription and its listed textual parallels; the source manuscript’s original-text section supplies the quoted 殆 reading. A Japanese woodblock text is held as book one of Hyōdai Bukei Shichisho Zenbun, books 1–7, corrected by Wang Huan (Osaka: Bunkindō, late Edo period). Waseda University Library, call number 文庫31 E1920.